Improving the diagnosis and prediction of customer churn: A heterogeneous hazard modeling approach

For your marketing

Possible founder application

Use it to frame a lifecycle hypothesis and define customer-level outcomes; compare its market and retention horizon with yours.

Reported result

What the study found

Adding heterogeneity significantly improves churn prediction. Churn is significantly linked to customer-service experience, failure-recovery efforts, and payment equity, with important differences across latent classes.

Research question

What the authors investigated

Can a heterogeneous hazard model improve diagnosis and prediction of customer churn, and how are churn rates related to service experience, recovery, and payment equity?

Design

Latent-class Weibull hazard model with time-varying covariates, allowing heterogeneity in baseline churn hazards and response parameters.

Sample and setting

Customer-level churn data from a continuous-service provider: a direct-to-home satellite television firm in a South American country.

Scope and transfer

Limitations to keep in view

Not stated on the publisher article-preview page fetched in this run.

Research sources

The publisher abstract explicitly identifies the latent-class Weibull/time-varying-covariate approach, the South American direct-to-home satellite-TV setting, improved prediction from heterogeneity, and the significant service, recovery, and payment-equity links. Direct publisher-page retrieval was recorded on 2026-10-09; profile summarizes the accessible record and does not claim independent review of inaccessible full text.